Testing mega-issue #24

New Issue

david · 2023-01-22T21:36:05-07:00

david commented

2023-01-22 21:36:05 -07:00

This will contain my general test plan as well as some more major test results.

First off, I want to do as much debugging of my RTL through simulation testing, and want to do that through more formalized regression tests as much as possible. As well I hope to have module level unit tests where it makes sense. (Potentially I could also make some test firmwares to aid this).

Obviously hardware bring-up will be its own thing, and I will want some test firmwares so I can do that whole process. Although, I can likely just run my regular firmware with some minor tweaks to do this.

Once basic hardware bringup is done (and I do any necessary board revisions :/), then comes the fun part, where I do full integration testing on the actual board.

I will start with a single channel on the 4-channel design, so I can use our 4th hydrophone and not deal with re-terminating the BNC connectors until later. I should be able to do the majority of my debugging here. The only major worry I have with moving to 4 channels when we add them is doing any automatic gain control. Although I'm not sure we need to do that to a significant amount. It can also be done on the software side, it doesn't need to happen in the firmware, as long as I export the gain settings.

This will contain my general test plan as well as some more major test results. First off, I want to do as much debugging of my RTL through simulation testing, and want to do that through more formalized regression tests as much as possible. As well I hope to have module level unit tests where it makes sense. (Potentially I could also make some test firmwares to aid this). Obviously hardware bring-up will be its own thing, and I will want some test firmwares so I can do that whole process. Although, I can likely just run my regular firmware with some minor tweaks to do this. Once basic hardware bringup is done (and I do any necessary board revisions :/), then comes the fun part, where I do full integration testing on the actual board. I will start with a single channel on the 4-channel design, so I can use our 4th hydrophone and not deal with re-terminating the BNC connectors until later. I should be able to do the majority of my debugging here. The only major worry I have with moving to 4 channels when we add them is doing any automatic gain control. Although I'm not sure we need to do that to a significant amount. It can also be done on the software side, it doesn't need to happen in the firmware, as long as I export the gain settings.

david commented

2023-01-22 21:37:55 -07:00

The first test that passes is I have blinky!

Today I fixed the issues with my SoC (most issues due to me not understanding the wishbone utilities I was using from amaranth-soc, and then some dumb addressing things), and got to a simple blinky.

I was able to debug all of my issues with simulation, pretty effectively, which gives me hope that the rest of the design will also go smoothly. But I am at least able to reliably debug my issues in sim, which will help a lot as behaviour gets more complex.

The first test that passes is I have blinky! Today I fixed the issues with my SoC (most issues due to me not understanding the wishbone utilities I was using from amaranth-soc, and then some dumb addressing things), and got to a simple blinky. I was able to debug all of my issues with simulation, pretty effectively, which gives me hope that the rest of the design will also go smoothly. But I am at least able to reliably debug my issues in sim, which will help a lot as behaviour gets more complex.

david commented

2023-01-29 16:58:31 -07:00

I've decided to go with unittest for my main test infrastructure. I've used it before, very simple test setup. Just need to flesh out a simple base class with amaranth-specific helpers for running simulation.

Sign in to join this conversation.